Arsenic Accumulation in a Paddy Field in ... - ACS Publications

9 downloads 0 Views 307KB Size Report
Mar 17, 2010 - and losses of soil As in a rice paddy field during irrigation and monsoon ...... (36) Ahmed, M. F.; Ahuja, S.; Alauddin, M.; Hug, S. J.; Lloyd, J. R.;.
Environ. Sci. Technol. 2010, 44, 2925–2931

Arsenic Accumulation in a Paddy Field in Bangladesh: Seasonal Dynamics and Trends over a Three-Year Monitoring Period JESSICA DITTMAR,† A N D R E A S V O E G E L I N , * ,†,‡ L I N D A C . R O B E R T S , †,‡ S T E P H A N J . H U G , ‡ GANESH C. SAHA,| M. ASHRAF ALI,§ A. BORHAN M. BADRUZZAMAN,§ AND RUBEN KRETZSCHMAR† Institute of Biogeochemistry and Pollutant Dynamics, Department of Environmental Sciences, ETH Zurich, CHN F23.2, CH-8092 Zurich, Switzerland, Eawag, Swiss Federal Institute of Aquatic Science and Technology, Ueberlandstrasse 133, CH-8600 Duebendorf, Switzerland, Department of Civil Engineering, Bangladesh University of Engineering and Technology, Dhaka 1000, Bangladesh, and Department of Civil Engineering, Dhaka University of Engineering and Technology, Dhaka 1700, Bangladesh

Received October 13, 2009. Revised manuscript received February 24, 2010. Accepted February 26, 2010.

Shallow groundwater, often rich in arsenic (As), is widely used for irrigation of dry season boro rice in Bangladesh. In the long term, this may lead to increasing As contents in rice paddy soils, which threatens rice yields, food quality, and human health. The objective of this study was to quantify gains and losses of soil As in a rice paddy field during irrigation and monsoon flooding over a three-year period. Samples were collected twice a year on a 3D-sampling grid to account for the spatially heterogeneous As distribution within the soil. Gains and losses of soil As in different depth segments were calculated using a mass-balance approach. Annual As input with irrigation water was estimated as 4.4 ( 0.4 kg ha-1 a-1. Within the top 40 cm of soil, the mean As accumulation over three years amounted to 2.4 ( 0.4 kg ha-1 a-1, implying that on average 2.0 kg ha-1 a-1 were lost from the soil. Seasonal changes of soil As showed that 1.05 to 2.1 kg ha-1 a-1 were lost during monsoon flooding. The remaining As-loss (up to 0.95 kg ha-1 a-1) was attributed to downward flow with percolating irrigation water. Despite these losses, we estimate that total As within the top 40 cm of soil at our field site would further increase by a factor of 1.5 to 2 by the year 2050 under current cultivation practices.

Introduction Rice production in Bangladesh was increased over the last decades to meet the rising food demand of the growing population. This was mainly achieved by increasing the cultivation of dry season boro rice (1), which provides higher * Corresponding author phone: +41-44-823 54 70; fax: +41-44823 52 10; e-mail: [email protected]. † ETH Zurich. ‡ Eawag. | Dhaka University of Engineering and Technology. § Bangladesh University of Engineering and Technology. 10.1021/es903117r

 2010 American Chemical Society

Published on Web 03/17/2010

yields and increased yield security compared to traditional varieties of wet season rice (2). Boro rice requires large amounts of water for irrigation, which is mainly extracted from shallow aquifers using tube wells (2). The groundwater extracted from shallow depth often contains high As concentrations (3, 4). Based on estimates of the amount of As extracted through shallow irrigation wells used for boro irrigation and the corresponding irrigated land area (5), on average ∼0.4 kg As ha-1 a-1 are introduced into boro paddy soils in Bangladesh. Arsenic may adversely affect rice yield and quality. Elevated As levels in rice straw and grain may lead to phytotoxicity. The transfer of As from rice straw into grain and the resulting human As uptake via rice represents an additional As uptake pathway in human nutrition, which may even dominate where human As intake via drinking water is limited (by provision of low-As drinking water) (6). In addition, increasing soil and plant As concentrations may affect human nutrition by decreasing the content of important nutrients such as Se, Ni, and Zn in rice grain (7). Numerous field studies reported relations between As in soil/irrigation water and rice plants (8-14). Pot experiments with natural or artificial soil/irrigation water As concentrations (8, 10, 14-18) revealed similar trends and yield reduction was observed to occur at high As concentrations (15, 17-19). First evidence of substantial yield reduction for rice grown on a paddy field in Bangladesh irrigated with As-rich groundwater, where As was strongly accumulated in the soil, was recently published (12, 14). Several field studies conducted in Bangladesh showed that irrigation with As-rich groundwater causes elevated As concentrations in the upper soil layers (11-13, 20-25). Panaullah et al. (12) investigated paddy fields in Bangladesh that are not deeply flooded during the monsoon season. They found that nearly all As added through irrigation was retained in the soil. In contrast, As content in topsoil of paddy fields affected by pronounced monsoon flooding decreased significantly over the monsoon season (22, 23, 25). During flood irrigation, arsenic is removed from spreading irrigation water by coprecipitation with Fe(III) and sorption to soil particles (26), which leads to laterally heterogeneous As inputs to paddy soils, particularly if the irrigation inlet remains unchanged for several years (12, 13, 23). Soil monitoring of As concentrations over one year and simple mass balance estimates after 15 years of irrigation indicated that this pattern of soil As accumulation persisted (23), despite the loss of As from the upper soil layers during monsoon flooding (22, 23, 25). In a recent study, we identified As release into floodwater followed by lateral removal toward rivers as a major pathway attenuating As accumulation in paddy soil (27). Calculations based on concentration gradients in the soil porewater and in monsoon floodwater suggested that 0.5-2.5 kg As ha-1 a-1 were released from the soil during monsoon flooding, corresponding to 13-62% of the estimated annual As input via irrigation of 4 kg ha-1 a-1. This clearly demonstrated that monsoon flooding is a key factor controlling the extent of long-term As accumulation in paddy soils. However, the uncertainties in these quantitative estimates are rather large and need to be further constrained. Additionally, other pathways such as As losses by leaching during irrigation must also be considered to obtain a comprehensive picture of As cycling during irrigation and monsoon seasons and to predict the possible long-term risks of As accumulation in soils caused by irrigation with As-rich groundwater. The objective of the present study was to derive a mass balance of As inputs and losses in a paddy field in Bangladesh irrigated with As-rich groundwater by determining the total VOL. 44, NO. 8, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2925

FIGURE 1. Map of the investigated field showing the sampling points (numbered in gray) monitored from December 2004 to December 2007. Large black numbers indicate the difference in soil As concentration (in mg kg-1) between December 2004 and December 2007. A map of the entire field site was previously published (23, 26). As mass within the top 40 cm of soil and its variations during irrigation and monsoon flooding over a 3-year monitoring period. The lateral and vertical distribution of soil As (23) were taken into account by collecting data on a three-dimensional sampling grid. The estimated As pools and fluxes and potential environmental implications are discussed based on long-term trend scenarios.

Materials and Methods Description of the Field Site. The study site is located in Munshiganj district 5 km north of the Ganges River and 1 km east of the Ichamati River. Since the early 1990s, the site was exclusively used for boro rice cultivation in the dry season. During rice growth from December until May, the fields are irrigated by a single shallow irrigation well extracting As-rich groundwater from 30 to 60 m depth. The well water is anoxic, has neutral pH, and contains 397 ( 7 µg L-1 As (∼84% As(III)), 2 mg L-1 P, and 11 mg L-1 Fe (26). For irrigation, an opening (inlet) is trenched into the field bund through which the water flows from the irrigation channel into the topographically flat field. The amount of irrigation water annually applied to the paddy fields at the study site may vary and several values have been reported in previous studies: 900 mm a-1 (22) and 960 mm a-1 (28) for the year 2004; 1200 mm a-1 for 2006 (29) and 1350 mm a-1 for 2007 (29). Calculations in this study were based on the mean of the four reported values ((standard error), that is, 1100 ( 100 mm a-1. The resulting average annual As input to the fields is 4.4 ( 0.4 kg As ha-1 a-1 ((standard error). During the monsoon, between mid June and late October, the area is flooded with floodwater levels reaching up to 4.5 m above soil surface (27, 28). The soil is tilled once a year by puddling before transplantation of the rice seedlings in December. Detailed descriptions of the soil properties at the site (23) and the nature of the underlying sediments (30) have previously been published. Soil Sampling. Soil samples were collected before soil puddling and rice transplantation in December 2004, 2005, 2006, and 2007 and after rice harvest in May 2005, 2006, and 2007. Figure 1 shows the investigated paddy field (2778 m2) sampled on 30 sampling points on a grid of 10 × 10 m with an additional eight samples on a finer 5 × 5 m grid (n ) 38 in total). Soil cores were collected either by pushing PVC tubes (diameter: 3.75 cm, length: 75 cm) into the soil (December 2004 and December 2006) or using a core sampler with PVC sleeves (diameter: 3.5 cm, length: 42 cm, Humax, Switzerland) (May 2005, December 2005, May 2006, May 2007, 2926

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 8, 2010

and December 2007). In December 2004, cores were divided into two depth segments (0-10 and 30-40 cm), and in December 2006 into three segments (0-10, 10-25, and 25-40 cm). At all other sampling times, soil samples were collected with higher vertical resolution (0-5, 5-10, 10-25, and 25-40 cm). The depth segments 0-10 and 10-25 cm were chosen as they represent the upper and lower portion of the soil above and including the plow pan, that is, the main rooting zone of rice. The soil segment 25-40 cm represents the unconsolidated alluvial material below the plow pan. The inlet position of the investigated field was maintained at the same place until 2006, but was shifted 15.3 m to the left in 2007 due to changes in land tenancy (Figure 1). Preparation and Analysis of Soil Samples. Soil samples were oven-dried at 60 °C, ground to